INTRODUCTION: WHAT IS PSYCHOLOGY?mood and alertness before the intervention began (at baseline) and
then again at the study end – the outcome. Many studies also repeat
their measures at several sub sequent time points, known as the
six-month, one-year (or whatever the delay is) follow-up. Studies that
take repeated measures over time are known as longitudinal, whereas
those that take an isolated snapshot are cross-sectional. The former
are considered superior because they can establish that a change in
a given factor at one time point has led to a change in another factor
later on. Cross-sectional studies, by contrast, are unable to demon-
strate the direction of causality between factors.
To judge whether any differences between baseline and outcome
measures, or between experimental and control groups, are significant,
psychologists use statistical tests. Generally these show the possibility
of the current outcome (or one even more extreme) being observed
if the null hypothesis were true. The null hypothesis is the opposite
of the hypothesis. In Langer and Rodin’s study, the hypothesis would
have been something like: “greater responsibilities and freedoms are
good for nursing home residents’ mental and physical health” whereas
the null hypothesis would have been something like: “greater respon-
sibilities and freedoms make no difference to nursing home residents’
mental and physical health.”
Psychologists typically consider an observed difference to be
statistically significant if there is a less than five percent chance that
it, or an even more extreme outcome, could have occurred if the null
hypothesis were true. Increasingly, researchers in psychology believe it
is also important to provide some measure of the size of the difference
that is observed, not just whether the difference is statistically signifi-
cant or not. This is known as the effect size, and in Langer and Rodin’s
study it could influence whether the difference in outcomes between
the experimental and control groups was clinically significant (that
is big enough to make the difference between a person meeting the
criteria, or not, for a diagnosis, such as for depression) or practically
meaningful, in terms of quality of life and increased longevity.
Finally, when Langer and Rodin came to publish their results
they would have had to submit them to a peer-reviewed academic
journal (they chose the Journal of Personality and Social Psychology).
Peer-review involves anonymous experts in the field scrutinizing the
study’s methodology and statistics to ensure the conclusions are
justified, and that there is enough information for other scientists to
replicate or extend the study should they wish to. Peer-reviewers can
reject or accept a paper, and in either case they will usually provide
useful feedback to the study authors. Nearly all the experiments
referred to throughout this book have been published in peer-
reviewed journals.